Artificial Intelligence or machine intelligence is intelligence demonstrated by machines. This means that machines are built and programmed to perform intelligent tasks such as understanding human speech, competing in strategic games, autonomously operating cars, etc which were activities exclusively performed by humans. In recent years artificial intelligence has become even more capable than earlier versions and have become advanced enough to be integrated into human activities such as traffic control, autonomous vehicles, high security operations etc. The traditional goals of AI include, reasoning, knowledge representation, planning, learning, natural language processing, perception and the ability to move and manipulate objects. Many tools are used in AI including versions of search and mathematical optimization, artificial neural networks, and methods based on statistics, probability and economics.
A typical AI machine analyzes it’s environment and takes actions that maximizes it’s chances of success. AI often revolves around the use of algorithms. An algorithm is a set of unambiguous instructions that a mechanical computer can execute. A complex algorithm is often built on top of other simpler algorithms. Many AI algorithms are capable of learning from data, they can enhance themselves by learning new strategies or rules that have worked well in the past or can write other algorithms themselves. An example of how AI works can be used to find the shortest possible routes from one location to the other on a wide map which contains houses, trees and other built up structures which in theory could make it very difficult in locating such routes. Matching several structures and other information to find out the shortest route from one city to the other within the shortest possible time could be quite difficult for humans, but AI is built to solve such problems within the shortest time possible to avoid the phenomenon called combinatorial explosion where the amount of time needed to solve a problem grows exponentially. The AI basically identifies and avoids a broad range of information that are not considered beneficial in solving the problem, thus enabling quicker solutions in record time.
Artificial Intelligence has developed a number of tools to solve the most difficult problems in computer science. These tools are;
SEARCH AND OPTIMIZATION
Many problems in AI can be solved theoretically by intelligently searching through many possible solutions; reasoning can be reduced to perform a search, so is planning algorithms and robotic algorithms. Many learning algorithms use search algorithms based on optimization. Simple exhaustive searches are rarely sufficient for most real world problems, the search space quickly grows to astronomical numbers. A very different kind of search came up in the 1990s based on the mathematical theory of optimization. This involves beginning a search incrementally with refinements until a stage is reached where no more refinements are made. Evolutionary computation which uses search optimization begins with a set of organisms which are allowed to mutate and recombine until only the fittest survive.
Logic is used for knowledge representation and problem solving, but it can be applied to other problems as well. Different forms of logic are used in AI research. Propositional logic involves truth functions such as “or” and “not”. Fuzzy logic is successfully used in control systems to allow experts to contribute vague rules such as increasing brake pressure in a train when it is very close to the station
PROBABILISTIC METHODS FOR UNCERTAIN REASONING
Many problems in AI require the agent to operate with incomplete or uncertain information. AI researchers have developed a number of powerful tools to solve these problems using methods from probability theory and economics. Bayesian networks are the very general tool used for various problems; reasoning using the Bayesian infererence algorithm, learning using the expectation – maximization algorithm, planning using decision networks, and perception using dynamic Bayesian networks. A key concept from the science of economics is intelligence, is utility, a measure of how valuable something is to an intelligent agent. Precise mathematical tools have been developed that analyze how an agent can make choices and plan, using decision theory, decision analysis and information value theory.
CLASSIFIERS AND STATISTICAL LEARNING METHODS.
The simplest AI applications can be divided into two. These are; classifiers and controllers. Controllers still classify conditions before inferring actions. Classifiers use pattern matching to determine a closest match. They can be tuned according to examples, making them very attractive for use in AI. A classifier can be trained in various ways, there are many statistical and machine learning approaches. The decision tree is perhaps the most widely used machine learning algorithm. Other widely used classifiers are the neural network, k- nearest neighbor algorithm, support vector machine and naive Bayes classifier. Classifier performance depends greatly on the characteristics of the data to be classified, such as the dataset size, distribution of samples across classes, the dimensionality and the level of noise.